You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Feature request summary
I'd love to see totalExecutionTime available as an assertion/budget target in Lightwallet.
What is the motivation or use case for changing this?
The metric values can exhibit high variance in certain situations that make it difficult to enable commit-gating. Total CPU execution time is a good indicator of overall performance characteristics that is much less variable according to our internal research projects (see #6775). It could be a nice middle ground between metric assertions and resource assertions.
How is this beneficial to Lighthouse?
Provides users with critical warnings on CPU performance changes that would otherwise go unnoticed.
The text was updated successfully, but these errors were encountered:
totalExecutionTime between navStart and TTI? End of trace?
This will be indeed be thorny because we don't wait for true TTI conditions in simulation, so I'll say navStart to endOfTrace to avoid any misrepresentation which will mostly align with navStart to TTI.
Of course there are other considerations like what we include (is it just JS execution time? probably not. should we split out JS time separately as a target? probably.). It could also be broken down by third party vs. first party, subject to the same device profiles, etc.
Just wanted to start a conversation around this :)
Feature request summary
I'd love to see
totalExecutionTime
available as an assertion/budget target in Lightwallet.What is the motivation or use case for changing this?
The metric values can exhibit high variance in certain situations that make it difficult to enable commit-gating. Total CPU execution time is a good indicator of overall performance characteristics that is much less variable according to our internal research projects (see #6775). It could be a nice middle ground between metric assertions and resource assertions.
How is this beneficial to Lighthouse?
Provides users with critical warnings on CPU performance changes that would otherwise go unnoticed.
The text was updated successfully, but these errors were encountered: